Hi Peter, On Sun, Sep 18, 2011 at 10:47:51PM +0800, Wu Fengguang wrote: > > BTW, I also compared the IO-less patchset and the vanilla kernel's > > JBOD performance. Basically, the performance is lightly improved > > under large memory, and reduced a lot in small memory servers. > > > > vanillla IO-less > > -------------------------------------------------------------------------------- > [...] > > 26508063 17706200 -33.2% JBOD-10HDD-thresh=100M/xfs-100dd-1M-16p-5895M-100M > > 23767810 23374918 -1.7% JBOD-10HDD-thresh=100M/xfs-10dd-1M-16p-5895M-100M > > 28032891 20659278 -26.3% JBOD-10HDD-thresh=100M/xfs-1dd-1M-16p-5895M-100M > > 26049973 22517497 -13.6% JBOD-10HDD-thresh=100M/xfs-2dd-1M-16p-5895M-100M > > > > There are still some itches in JBOD.. > > OK, in the dirty_bytes=100M case, I find that the bdi threshold _and_ > writeout bandwidth may drop close to 0 in long periods. This change > may avoid one bdi being stuck: > > /* > * bdi reserve area, safeguard against dirty pool underrun and disk idle > * > * It may push the desired control point of global dirty pages higher > * than setpoint. It's not necessary in single-bdi case because a > * minimal pool of @freerun dirty pages will already be guaranteed. > */ > - x_intercept = min(write_bw, freerun); > + x_intercept = min(write_bw + MIN_WRITEBACK_PAGES, freerun); After lots of experiments, I end up with this bdi reserve point + x_intercept = bdi_thresh / 2 + MIN_WRITEBACK_PAGES; together with this chunk to avoid a bdi stuck in bdi_thresh=0 state: @@ -590,6 +590,7 @@ static unsigned long bdi_position_ratio( */ if (unlikely(bdi_thresh > thresh)) bdi_thresh = thresh; + bdi_thresh = max(bdi_thresh, (limit - dirty) / 8); /* * scale global setpoint to bdi's: * bdi_setpoint = setpoint * bdi_thresh / thresh The above changes are good enough to keep reasonable amount of bdi dirty pages, so the bdi underrun flag ("[PATCH 11/18] block: add bdi flag to indicate risk of io queue underrun") is dropped. I also tried various bdi freerun patches, however the results are not satisfactory. Basically the bdi reserve area approach (this patch) yields noticeably more smooth/resilient behavior than the freerun/underrun approaches. I noticed that the bdi underrun flag could lead to sudden surge of dirty pages (especially if not safeguarded by the dirty_exceeded condition) in the very small window.. To dig performance increases/drops out of the large number of test results, I wrote a convenient script (attached) to compare the vmstat:nr_written numbers between 2+ set of test runs. It helped a lot for fine tuning the parameters for different cases. The current JBOD performance numbers are encouraging: $ ./compare.rb JBOD*/*-vanilla+ JBOD*/*-bgthresh3+ 3.1.0-rc4-vanilla+ 3.1.0-rc4-bgthresh3+ ------------------------ ------------------------ 52934365 +3.2% 54643527 JBOD-10HDD-thresh=100M/ext4-100dd-1M-24p-16384M-100M:10-X 45488896 +18.2% 53785605 JBOD-10HDD-thresh=100M/ext4-10dd-1M-24p-16384M-100M:10-X 47217534 +12.2% 53001031 JBOD-10HDD-thresh=100M/ext4-1dd-1M-24p-16384M-100M:10-X 32286924 +25.4% 40492312 JBOD-10HDD-thresh=100M/xfs-10dd-1M-24p-16384M-100M:10-X 38676965 +14.2% 44177606 JBOD-10HDD-thresh=100M/xfs-1dd-1M-24p-16384M-100M:10-X 59662173 +11.1% 66269621 JBOD-10HDD-thresh=800M/ext4-10dd-1M-24p-16384M-800M:10-X 57510438 +2.3% 58855181 JBOD-10HDD-thresh=800M/ext4-1dd-1M-24p-16384M-800M:10-X 63691922 +64.0% 104460352 JBOD-10HDD-thresh=800M/xfs-100dd-1M-24p-16384M-800M:10-X 51978567 +16.0% 60298210 JBOD-10HDD-thresh=800M/xfs-10dd-1M-24p-16384M-800M:10-X 47641062 +6.4% 50681038 JBOD-10HDD-thresh=800M/xfs-1dd-1M-24p-16384M-800M:10-X The common single disk cases also see good numbers except for slight drops in the dirty_bytes=100MB case: $ ./compare.rb thresh*/*vanilla+ thresh*/*bgthresh3+ 3.1.0-rc4-vanilla+ 3.1.0-rc4-bgthresh3+ ------------------------ ------------------------ 4092719 -2.5% 3988742 thresh=100M/ext4-10dd-4k-8p-4096M-100M:10-X 4956323 -4.0% 4758884 thresh=100M/ext4-1dd-4k-8p-4096M-100M:10-X 4640118 -0.4% 4621240 thresh=100M/ext4-2dd-4k-8p-4096M-100M:10-X 3545136 -3.5% 3420717 thresh=100M/xfs-10dd-4k-8p-4096M-100M:10-X 4399437 -0.9% 4361830 thresh=100M/xfs-1dd-4k-8p-4096M-100M:10-X 4100655 -3.3% 3964043 thresh=100M/xfs-2dd-4k-8p-4096M-100M:10-X 4780624 -0.1% 4776216 thresh=1G/ext4-10dd-4k-8p-4096M-1024M:10-X 4904565 +0.0% 4905293 thresh=1G/ext4-1dd-4k-8p-4096M-1024M:10-X 3578539 +9.1% 3903390 thresh=1G/xfs-10dd-4k-8p-4096M-1024M:10-X 4029890 +0.8% 4063717 thresh=1G/xfs-1dd-4k-8p-4096M-1024M:10-X 2449031 +20.0% 2937926 thresh=1M/ext4-10dd-4k-8p-4096M-1M:10-X 4161896 +7.5% 4472552 thresh=1M/ext4-1dd-4k-8p-4096M-1M:10-X 3437787 +18.8% 4085707 thresh=1M/ext4-2dd-4k-8p-4096M-1M:10-X 1921914 +14.8% 2206897 thresh=1M/xfs-10dd-4k-8p-4096M-1M:10-X 2537481 +65.8% 4207336 thresh=1M/xfs-1dd-4k-8p-4096M-1M:10-X 3329176 +12.3% 3739888 thresh=1M/xfs-2dd-4k-8p-4096M-1M:10-X 4587856 +1.8% 4672501 thresh=400M-300M/ext4-10dd-4k-8p-4096M-400M:300M-X 4883525 +0.0% 4884957 thresh=400M-300M/ext4-1dd-4k-8p-4096M-400M:300M-X 4799105 +2.3% 4907525 thresh=400M-300M/ext4-2dd-4k-8p-4096M-400M:300M-X 3931315 +3.0% 4048277 thresh=400M-300M/xfs-10dd-4k-8p-4096M-400M:300M-X 4238389 +3.9% 4401927 thresh=400M-300M/xfs-1dd-4k-8p-4096M-400M:300M-X 4032798 +2.3% 4123838 thresh=400M-300M/xfs-2dd-4k-8p-4096M-400M:300M-X 2425253 +35.2% 3279302 thresh=8M/ext4-10dd-4k-8p-4096M-8M:10-X 4728506 +2.2% 4834878 thresh=8M/ext4-1dd-4k-8p-4096M-8M:10-X 2782860 +62.1% 4511120 thresh=8M/ext4-2dd-4k-8p-4096M-8M:10-X 1966133 +24.3% 2443874 thresh=8M/xfs-10dd-4k-8p-4096M-8M:10-X 4238402 +1.7% 4308416 thresh=8M/xfs-1dd-4k-8p-4096M-8M:10-X 3299446 +13.3% 3739810 thresh=8M/xfs-2dd-4k-8p-4096M-8M:10-X Thanks, Fengguang